OpenPose: Optimized for Mobile Deployment
Human pose estimation
OpenPose is a machine learning model that estimates body and hand pose in an image and returns location and confidence for each of 19 joints.
This model is an implementation of OpenPose found here.
More details on model performance across various devices, can be found here.
Model Details
- Model Type: Pose estimation
- Model Stats:
- Model checkpoint: body_pose_model.pth
- Input resolution: 240x320
- Number of parameters: 52.3M
- Model size: 200 MB
| Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model |
|---|---|---|---|---|---|---|---|---|
| OpenPose | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 11.495 ms | 0 - 917 MB | FP16 | NPU | -- |
| OpenPose | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 11.517 ms | 1 - 3 MB | FP16 | NPU | -- |
| OpenPose | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 11.59 ms | 0 - 226 MB | FP16 | NPU | -- |
| OpenPose | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 8.562 ms | 0 - 136 MB | FP16 | NPU | -- |
| OpenPose | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 8.597 ms | 1 - 18 MB | FP16 | NPU | -- |
| OpenPose | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 8.799 ms | 1 - 32 MB | FP16 | NPU | -- |
| OpenPose | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 8.892 ms | 0 - 25 MB | FP16 | NPU | -- |
| OpenPose | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 8.596 ms | 0 - 18 MB | FP16 | NPU | -- |
| OpenPose | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 8.741 ms | 0 - 23 MB | FP16 | NPU | -- |
| OpenPose | SA7255P ADP | SA7255P | TFLITE | 769.55 ms | 0 - 19 MB | FP16 | NPU | -- |
| OpenPose | SA7255P ADP | SA7255P | QNN | 769.528 ms | 1 - 10 MB | FP16 | NPU | -- |
| OpenPose | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 11.558 ms | 0 - 907 MB | FP16 | NPU | -- |
| OpenPose | SA8255 (Proxy) | SA8255P Proxy | QNN | 11.479 ms | 1 - 4 MB | FP16 | NPU | -- |
| OpenPose | SA8295P ADP | SA8295P | TFLITE | 26.259 ms | 0 - 22 MB | FP16 | NPU | -- |
| OpenPose | SA8295P ADP | SA8295P | QNN | 25.345 ms | 1 - 19 MB | FP16 | NPU | -- |
| OpenPose | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 11.525 ms | 0 - 866 MB | FP16 | NPU | -- |
| OpenPose | SA8650 (Proxy) | SA8650P Proxy | QNN | 11.512 ms | 1 - 3 MB | FP16 | NPU | -- |
| OpenPose | SA8775P ADP | SA8775P | TFLITE | 29.082 ms | 0 - 19 MB | FP16 | NPU | -- |
| OpenPose | SA8775P ADP | SA8775P | QNN | 28.874 ms | 1 - 10 MB | FP16 | NPU | -- |
| OpenPose | QCS8275 (Proxy) | QCS8275 Proxy | TFLITE | 769.55 ms | 0 - 19 MB | FP16 | NPU | -- |
| OpenPose | QCS8275 (Proxy) | QCS8275 Proxy | QNN | 769.528 ms | 1 - 10 MB | FP16 | NPU | -- |
| OpenPose | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 11.519 ms | 0 - 908 MB | FP16 | NPU | -- |
| OpenPose | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 11.468 ms | 1 - 4 MB | FP16 | NPU | -- |
| OpenPose | QCS9075 (Proxy) | QCS9075 Proxy | TFLITE | 29.082 ms | 0 - 19 MB | FP16 | NPU | -- |
| OpenPose | QCS9075 (Proxy) | QCS9075 Proxy | QNN | 28.874 ms | 1 - 10 MB | FP16 | NPU | -- |
| OpenPose | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 23.431 ms | 0 - 135 MB | FP16 | NPU | -- |
| OpenPose | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 22.383 ms | 0 - 23 MB | FP16 | NPU | -- |
| OpenPose | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 11.982 ms | 1 - 1 MB | FP16 | NPU | -- |
| OpenPose | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 12.273 ms | 101 - 101 MB | FP16 | NPU | -- |
License
- The license for the original implementation of OpenPose can be found here.
- The license for the compiled assets for on-device deployment can be found here
References
- OpenPose: Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields
- Source Model Implementation
Community
- Join our AI Hub Slack community to collaborate, post questions and learn more about on-device AI.
- For questions or feedback please reach out to us.
Usage and Limitations
Model may not be used for or in connection with any of the following applications:
- Accessing essential private and public services and benefits;
- Administration of justice and democratic processes;
- Assessing or recognizing the emotional state of a person;
- Biometric and biometrics-based systems, including categorization of persons based on sensitive characteristics;
- Education and vocational training;
- Employment and workers management;
- Exploitation of the vulnerabilities of persons resulting in harmful behavior;
- General purpose social scoring;
- Law enforcement;
- Management and operation of critical infrastructure;
- Migration, asylum and border control management;
- Predictive policing;
- Real-time remote biometric identification in public spaces;
- Recommender systems of social media platforms;
- Scraping of facial images (from the internet or otherwise); and/or
- Subliminal manipulation
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
🙋
Ask for provider support
HF Inference deployability: The HF Inference API does not support keypoint-detection models for pytorch
library.